Why Not Antivirals? Reconsidering Influenza Treatment in Children

Abstract

In this issue of Pediatrics, Huang et al1 shed important light on the association between antiviral use and reduced influenza-related hospitalizations in children. Using a large national health care database from Taiwan and accounting for a number of clinically relevant factors, the case-control study decisively showed that early antiviral use within 48 hours of symptoms onset was associated with an 81% reduction in the risk of influenza-related hospitalization. Rigorous sensitivity analyses challenging study assumptions did not alter their conclusions. These findings closely resemble several other well performed observational studies powered to assess the outcome of influenza hospitalization and antiviral use.2–12 These studies varied in their definition of influenza-related or any cause hospitalization and whether antivirals were used early or late. Several of these studies also controlled for vaccination status, duration of symptoms, and health care use and assessed negative controls for residual biases. Performed over different influenza seasons in various countries using different methods and data sources, they clearly demonstrate clinically impactful reductions in hospitalization and influenza-related complications with antiviral use.Randomized controlled trials (RCTs) are the gold standard study design for assessing therapeutic effectiveness, but they have limitations, particularly when the study outcome is an uncommon event, such as influenza-related hospitalization. In contrast, observational studies use much larger data from real-world settings to provide an adequate sample size to evaluate uncommon occurrences. When rigorously conducted, observational studies can provide important evidence when RCTs cannot.13,14 In fact, the US Food and Drug Administration (FDA) recently stressed the power of observational studies, stating that nonrandomized data can lead to “actionable causal conclusions.”15 The Huang study published this month in Pediatrics, put in the context of other studies in outpatients (Figure 1), provides “actionable causal conclusions” that early antiviral treatment shortly after symptom onset decreases influenza-associated hospitalizations and associated complications.Despite repeated studies showing the clinical benefit of antiviral treatment in outpatients on hospitalization, why are they not used more? It is likely because of the controversy that has surrounded the association between antiviral use and reduced hospitalizations for almost 2 decades. This is largely because RCTs are underpowered to assess a hospitalization outcome.16–18 RCTs for influenza antivirals were performed to assess minor endpoints in outpatients with uncomplicated influenza, such as time to alleviation of symptoms, and were not designed to evaluate serious or uncommon outcomes. The antiviral oseltamivir was approved in 1999 based on 2 RCTs conducted in previously healthy adolescents and adults with laboratory-confirmed influenza.19,20 Patients treated within 48 hours of symptom onset randomized to oseltamivir had a 1.3 days reduction of illness when compared with placebo recipients.19,20 In 2003, several clinical trials were combined in a publication in Arch Intern Med demonstrating that oseltamivir treatment reduced all-cause hospitalization by 59% (P = 0.02) and lower respiratory tract infections treated with antibiotics by 55% (P = <0.001).21 In 2006, the Cochrane group published a meta-analysis of RCTs concluding a 68% reduction in pneumonia with oseltamivir treatment of influenza.22 However, this publication triggered widespread controversy, later suggesting that clinical trial data had been withheld by Roche (the manufacturer of oseltamivir) and that the oseltamivir benefit for reducing influenza-related hospitalization had been oversold.Additional trial data were eventually provided by Roche and another Cochrane review published in 2014 found that oseltamivir treatment of outpatients with influenza reduced pneumonia by 55% but did not reduce any cause hospitalization.23 Importantly, the Cochrane review analyzed the risk of hospitalization in the influenza-like illness population (rather than the laboratory-confirmed influenza population). In their primary analysis of oseltamivir and hospitalization, 33% of participants included did not have influenza, likely biasing results toward the null. The same Cochrane review also reported that oseltamivir treatment reduced the duration of influenza symptoms in children by 29 hours, a clinically meaningful reduction, similar to subsequent meta-analyses in children24 and data provided to support FDA approval. Yet, this Cochrane review led many to conclude that oseltamivir had minimal or no impact on hospitalizations.To address this controversy surrounding oseltamivir’s effect on pneumonia and hospitalization, the Multiparty Group for Advice on Science assembled a research team25 and obtained an unrestricted grant from Roche to perform an additional meta-analysis that included more completed trial information and individual patient data rather than aggregated results, the preferred analytic method for meta-analyses.26,27 This meta-analysis of oseltamivir treatment in outpatient adults with laboratory-confirmed uncomplicated influenza published by Dobson et al in the Lancet25 reported reductions in both hospitalization by 63% and pneumonia by 44%.25,27Several subsequent meta-analyses reported mixed findings for influenza antivirals and influenza-related hospitalization using clinical trial data.18,24,28,29 However, there were significant limitations to these studies. Outpatient RCTs of antivirals were not designed to evaluate severe outcomes such as influenza complications or hospitalization. In fact, pivotal clinical trials that led to FDA approval of oseltamivir did not include hospitalization as a primary or sometimes even a tertiary outcome. In clinical trials, especially older trials, nonprimary end points may be less precisely chosen, measured, analyzed, and reported compared with primary end points. Meta-analysis of RCTs in both children and adults are severely underpowered to evaluate hospitalization as an outcome.18,24 For example, a recent meta-analysis in JAMA Internal Medicine, one of the largest to date, included 6166 individuals, 3324 of whom received oseltamivir. The study concluded there was no reduction in any-cause hospitalization with oseltamivir (risk ratio [RR] 0.77; 95% CI 0.47–1.27). However, another valid interpretation would be that the study was substantially underpowered to determine a precise effect estimate and that oseltamivir may indeed reduce hospitalization.16,17 One of us (JA) recently reported that if the true RR for oseltamivir treated relative to untreated patients was 0.77 (a 23% reduction), a prospective trial would need approximately 87 000 participants (48 704 oseltamivir exposed and 38 963 unexposed) to reject the null hypothesis.16,17 Simply put, the question of “Does oseltamivir reduce hospitalization?” cannot be properly answered with pooled data from RCTs of oseltamivir treatment in outpatients conducted to date and must be assessed in observational studies, such as the Huang study in this issue of Pediatrics.It is biologically plausible that early use of influenza antivirals can reduce hospitalization for influenza-related illnesses because they reduce viral load and viral-mediated inflammation.30–32 RCTs and well-designed observational studies have also shown that antivirals reduce influenza complications such as pneumonia, otitis media, sinusitis, and neurologic events in children and adolescents.2,12,24,25,33 The Centers for Disease Control and Prevention (CDC) estimates that current antiviral prescribing practices prevent up to 14 184 hospitalizations per season and that a 50% improvement in access, testing, and prescribing could avert over 71 000 hospitalizations.34For nearly 20 years, the American Academy of Pediatrics (AAP), CDC, and the Infectious Disease Society of America (IDSA) have recommended prompt antiviral treatment in all high-risk children with influenza and consideration of use in those within 48 hours of symptom onset.35–37 However, despite these recommendations, pediatric providers remain very hesitant to prescribe antivirals and at the same time overprescribe antibiotics. Approximately 40% of high-risk children with influenza do not receive a recommended antiviral, a trend that has worsened since the COVID-19 pandemic.38,39 Recent findings published in Pediatrics40,41 on reasons for the underprescribing of antivirals for influenza are shockingly similar to results from a study conducted 15 years ago by the CDC42 and include the following: 1) lack of awareness of national guidelines, 2) questions about the effectiveness of antivirals, and 3) concern about adverse events associated with antivirals, particularly vomiting and neuropsychiatric events.Although we have addressed the first 2 of these issues, we would like to conclude this commentary discussing the safety of oseltamivir. Although over 2 decades of clinical studies support the safety of influenza antivirals, many providers remain very concerned about antiviral-related vomiting and neuropsychiatric events. However, vomiting is not a common adverse event with oseltamivir. Oseltamivir RCTs in children demonstrate a 5.8% increase in vomiting compared with placebo.24 This means that for every 17 patients treated with oseltamivir, 1 will experience additional vomiting. Oseltamivir-related nausea and vomiting can be reduced when taken with food and respond extremely well to ondansetron and other antiemetics and should not be the reason to withhold oseltamivir when indicated.18,30Neuropsychiatric events are also an area of a concern and were added to the oseltamivir label after case reports of such events were reported (often after a single dose of oseltamivir), mainly out of Japan.12 Early evaluations of these events used a self-controlled case series design, which cannot account for underlying influenza infection.43 It is now well established that the influenza virus can precipitate neurologic complications, including acute necrotizing encephalopathy with high mortality.44,45 These complications include a wide spectrum of neurologic events, from seizures or transient encephalopathy lasting a few hours to fulminant acute necrotizing encephalopathy with progression from uncomplicated febrile upper respiratory illness to death within 24 hours.44,45 Recently, one of us (JA) performed a retrospective cohort study of 692 295 children with Medicaid insurance in the state of Tennessee and found that oseltamivir treatment was associated with a 47% reduction in neuropsychiatric events resulting in hospitalization (incidence rate ratio 0.53; 95% CI 0.33–0.88). These findings were in line with other well controlled studies using different data sources and study designs showing that influenza itself is associated with neuropsychiatric events and that oseltamivir use reduces this complication.12,46–52Although oseltamivir is the most prescribed antiviral, there is slow but growing use of baloxavir in both children and adults. Baloxavir is a cap-dependent endonuclease inhibitor that is FDA approved for early treatment of uncomplicated influenza in those aged at least 5 years that is as or more effective than oseltamivir in shortening duration of influenza symptoms.31,32 Baloxavir has some therapeutic advantages over oseltamivir, including a greater and more rapid reduction in viral load, reduced household influenza transmission, single-dose administration, better efficacy for influenza B strains, and better palatability.31,32,53–58 Some studies suggest that baloxavir may have a greater impact on influenza complications and hospitalization than oseltamivir33,59,60 and does not have vomiting or neuropsychiatric events listed on the label. Most of the antiviral prescribing in the Huang study was for oseltamivir, which included only 3 baloxavir users. Although baloxavir has been shown to be cost-effective,53 it is more expensive than oseltamivir and not always covered by Medicaid or insurance, likely limiting its widespread use.With greater public hesitancy regarding vaccines, declining seasonal influenza vaccination rates, and recent influenza seasons with record pediatric deaths and hospitalizations, providers should strongly consider the use of antivirals in high-risk and hospitalized patients.44,61,62 Many studies support the CDC, AAP, and IDSA recommendations on the use of oseltamivir and other influenza antivirals and demonstrate they are cost-effective, especially in high-risk individuals53,63–65 Outside of hospitalization, benefits of antivirals also include reduced household transmission of influenza, decreased influenza complications, and faster return to school for patients and work for parents.2,12,24,25,31–33,53–58,66–70 The CDC, AAP, IDSA and influenza experts all recommend prompt antiviral treatment of influenza as soon as possible after symptom onset in persons who are at increased risk of influenza complications, including children aged under 2 years and those with high-risk co-morbidities. The time has come for antivirals to be used in children with influenza.